Click here to go to the forum index Click here for the home page
 
Author Message

<  TAP and patch development  ~  TimerCheck -- for spotting corruption

Page 3 of 4
Goto page Previous  1, 2, 3, 4  Next
R2-D2
Posted: Sat Dec 20, 2008 5:21 pm Reply with quote
Frequent contributor Joined: 18 Dec 2006 Posts: 12148
Ooh. Just found the one place where the TSID might get changed and put a monitor on it. Ran it for a while and nothing happened. I then changed over to not checking whether there was an alteration being made and, freakily, I got a hit on a channel change to Sky3. Unfortunately, it seems to have been a change over to the correct data, so I may have just missed the bad change. Sad Nothing new has happened for a while now, but that would fit with it being a rare issue.

Now, this is on the routine that processes the SDT section... Anyone remember that weird, very occasional problem with the last four bytes of the recorded SDT packets? Hmm...

_________________
Troubleshooting -- User Manual -- Dark Side of the Matrix: Firmwares and Patches
View user's profile Send private message Visit poster's website
chunkywizard
Posted: Sat Dec 20, 2008 5:27 pm Reply with quote
MyStuff Team Joined: 13 May 2005 Posts: 9319 Location: Sindlesham, Berkshire Tx: Hannington
Sounds very interesting. If it's rare can you write a monitor a few of us can run to find the issue more quickly?

CW

_________________
Toppy:TF5800 F/W: 1365PaTcHeD
TAPs: MyStuff | EPG2MEI | FontManager | SecCache UK | Extend | FastScanGUI | EIT_sub_game | MyInfo | ScreenCapture with OSD | MHEG On/Off
MyStuff Manual & Logos | Known Bugs | Recommended F/W
View user's profile Send private message Visit poster's website
R2-D2
Posted: Sat Dec 20, 2008 6:05 pm Reply with quote
Frequent contributor Joined: 18 Dec 2006 Posts: 12148
It's hard-coded to 5.13.40 at the moment, and requires a serial port connection, so not really practical. I've found a flag that does "don't update", so anyone who finds themselves getting lots of timer corruptions can maybe have a patch to try out...

_________________
Troubleshooting -- User Manual -- Dark Side of the Matrix: Firmwares and Patches
View user's profile Send private message Visit poster's website
Anthony
Posted: Sun Dec 21, 2008 9:32 am Reply with quote
Regular contributor Joined: 15 Feb 2006 Posts: 58
glenmcfar on 19/12/200 at 7:17pm wrote:
'Success - no bad timers found'

Thought I might be a candidate for this - I have merging turned off.
Saying that, I don't really see many problems in THIS department.

Glen.

glenmcfar on 20/12/2008 at 9:57am wrote:
Code:
-
- 2008-12-20 09:55:24 (0d 1h 2m 18s) -
 11) BROKEN: tnr=3 sat=0 net=233a ts=6000 svc=2100 Mon 22 Dec 18:58  62m  29 E4 Hollyoaks.rec
  should be: tnr=3 sat=0 net=233a ts=2011 svc=2100
 12) BROKEN: tnr=3 sat=0 net=233a ts=6000 svc=2100 Thu 25 Dec 18:58  34m  29 E4 Hollyoaks.rec
  should be: tnr=3 sat=0 net=233a ts=2011 svc=2100
 13) BROKEN: tnr=3 sat=0 net=233a ts=6000 svc=2100 Fri 26 Dec 18:58  34m  29 E4 Hollyoaks.rec
  should be: tnr=3 sat=0 net=233a ts=2011 svc=2100
 37) BROKEN: tnr=3 sat=0 net=233a ts=6000 svc=20fa Sun 21 Dec 20:58  64m  14 More 4 Grand Designs.rec
  should be: tnr=3 sat=0 net=233a ts=2011 svc=20fa
 38) BROKEN: tnr=3 sat=0 net=233a ts=6000 svc=20fa Mon 22 Dec 16:58  69m  14 More 4 Grand Designs.rec
  should be: tnr=3 sat=0 net=233a ts=2011 svc=20fa
 39) BROKEN: tnr=3 sat=0 net=233a ts=6000 svc=20fa Mon 22 Dec 21:58  69m  14 More 4 Grand Designs.rec
  should be: tnr=3 sat=0 net=233a ts=2011 svc=20fa


Found some!!!!!!!!!!


I wonder if these timers were not corrupt the first time Glen checked??

Kind regards

Anthony
View user's profile Send private message
R2-D2
Posted: Sun Dec 21, 2008 9:49 am Reply with quote
Frequent contributor Joined: 18 Dec 2006 Posts: 12148
Anthony wrote:
I wonder if these timers were not corrupt the first time Glen checked??
He's maybe been trying to provoke things, or just caught it at a (nicely) bad time. However, you're right if you're suggesting that false negatives might be seen if the service data is still wrong at the time of the check, although there's a good chance some timers (on those channels) are still present from when the data was correct.

_________________
Troubleshooting -- User Manual -- Dark Side of the Matrix: Firmwares and Patches
View user's profile Send private message Visit poster's website
Anthony
Posted: Sun Dec 21, 2008 9:59 am Reply with quote
Regular contributor Joined: 15 Feb 2006 Posts: 58
If I understand correctly the TSID is used to find the right Mux but is there any valid reason why the TSID would change for a LCN? like at 3am in the morning when the Tele shopping channels are broadcasting and most people have a watch timer setup

Kind regards

Anthony
View user's profile Send private message
R2-D2
Posted: Sun Dec 21, 2008 10:29 am Reply with quote
Frequent contributor Joined: 18 Dec 2006 Posts: 12148
Anthony wrote:
is there any valid reason why the TSID would change for a LCN?
Yes, because that's what the system is meant to cope with, according to the standard. In this case, it's the Actual SDT, which is for the currently tuned mux, so it's a bit odd for that to change. Smile Hmm... but there could easily be a timing issue, with the section data arriving just after the main channel has been changed, due to the delays in the internal message handling... that's something I can probably try to provoke...

_________________
Troubleshooting -- User Manual -- Dark Side of the Matrix: Firmwares and Patches
View user's profile Send private message Visit poster's website
glenmcfar
Posted: Sun Dec 21, 2008 10:41 am Reply with quote
Frequent contributor Joined: 07 Dec 2006 Posts: 4519 Location: Dundonald, Ayrshire, Scotland
R2-D2 wrote:
Anthony wrote:
I wonder if these timers were not corrupt the first time Glen checked??
He's maybe been trying to provoke things, or just caught it at a (nicely) bad time. However, you're right if you're suggesting that false negatives might be seen if the service data is still wrong at the time of the check, although there's a good chance some timers (on those channels) are still present from when the data was correct.


This is all above my heads. I was provided with a TAP, which I've been running every couple of hours when I'm in front of the TV. No provoking here, guv'n'r !

Glen.

_________________
H/W: TF5800 | URC-7555 | Asus | Best Firmware Ever!
A/S: SecCacheUK, EitSub, EPG2MEI, Display, Extend, QuickJump, FontManager, TapLauncher, MyStuff
T/L: TapCommander, Surfer, MeiSearch, MediaManager | HDFW, CutAds, Sudoku
View user's profile Send private message
R2-D2
Posted: Sun Dec 21, 2008 11:04 am Reply with quote
Frequent contributor Joined: 18 Dec 2006 Posts: 12148
glenmcfar wrote:
No provoking here, guv'n'r !
Might you be the sort of chap who would regularly change channel in rather quick succession? Something like <10s between changes?

_________________
Troubleshooting -- User Manual -- Dark Side of the Matrix: Firmwares and Patches
View user's profile Send private message Visit poster's website
glenmcfar
Posted: Sun Dec 21, 2008 1:37 pm Reply with quote
Frequent contributor Joined: 07 Dec 2006 Posts: 4519 Location: Dundonald, Ayrshire, Scotland
Not normally, no. The TV is usually tuned to CBeebies or a music channel and any browsing is done in the MyStuff EPG rather than channel hopping.

Of course, it MIGHT happen occasionally.

Are you on to something, R2?

Glen.

_________________
H/W: TF5800 | URC-7555 | Asus | Best Firmware Ever!
A/S: SecCacheUK, EitSub, EPG2MEI, Display, Extend, QuickJump, FontManager, TapLauncher, MyStuff
T/L: TapCommander, Surfer, MeiSearch, MediaManager | HDFW, CutAds, Sudoku
View user's profile Send private message
R2-D2
Posted: Sun Dec 21, 2008 4:54 pm Reply with quote
Frequent contributor Joined: 18 Dec 2006 Posts: 12148
glenmcfar wrote:
Are you on to something
As above, I've a suspicion that an SDT (Actual) section arrives and an interrupt happens, then the filter decides the system is interested in it and generates an internal message. And, in the microseconds before the message is acted upon (maybe more if the system is busy), the channel is changed and then this means the wrong data will be updated (since it won't match!). Seconds later the SDT (Actual) section arrives for the (new) current channel and is acted upon, correcting the data. In that (rare) gap when the data is wrong, if you happen to have a timer being updated for the new current channel, it will get the wrong service data, and this will be preserved until the next time it is updated. As noted previously, if you have MyStuff set up to remove padding across channels or to not merge recordings on the same channel then the chances of a timer update happening at the bad time is hugely increased.

Fixing this sort of thing is going to be a bit tricky, but if I had the source code I'd change the filter/handler/message to include the channel it was for, rather than choosing the one that is current at the time the handler does its thing. And it goes without saying that this is a fairly major conceptual mistake which could easily have been reproduced in plenty of other places in the firmware.

And a side note: this sort of thing is likely to make even correct timers exhibit the chase play problems, if they were started during that bad window.

(You know, sometimes I'm amazed it works at all! Smile)

_________________
Troubleshooting -- User Manual -- Dark Side of the Matrix: Firmwares and Patches
View user's profile Send private message Visit poster's website
ryclark
Posted: Sun Dec 21, 2008 6:20 pm Reply with quote
Frequent contributor Joined: 10 Jul 2005 Posts: 1604 Location: Shropshire
Well, I am certainly amazed that a Toppy ever worked at all without all your intervention Very Happy Thanks once again for all your efforts and investigations which make such interesting reading even if I don't quite follow all of it.

_________________
TF5800t, TS On, F/W: MS6 Recommended F/W 12/9/2009 -Fm
TAPs: PcControl B1.4; MyInfo B5.6; EPG2MEI v0.96; MyStuff 6.6; MHEG On/Off A3; Extend v1.7; SecCache (UK) v0.4; EIT Sub (Game) v0.6; TSSaver v0.5;
Sig generated by MyInfo on 13/10/14
?
View user's profile Send private message
R2-D2
Posted: Mon Dec 22, 2008 9:20 am Reply with quote
Frequent contributor Joined: 18 Dec 2006 Posts: 12148
If my suspicions are correct then I'd also expect the channel name to change momentarily, too. I think there'll normally be a window of about 10s before things get corrected. It's pretty unlikely, but has anyone seen this sort of thing before?

_________________
Troubleshooting -- User Manual -- Dark Side of the Matrix: Firmwares and Patches
View user's profile Send private message Visit poster's website
R2-D2
Posted: Mon Dec 22, 2008 10:57 am Reply with quote
Frequent contributor Joined: 18 Dec 2006 Posts: 12148
[Td] TimerData stops the SDT updating the service IDs. It'd probably take a long time to prove it works (or a shorter time to prove it doesn't...).

_________________
Troubleshooting -- User Manual -- Dark Side of the Matrix: Firmwares and Patches
View user's profile Send private message Visit poster's website
Wallace
Posted: Mon Dec 22, 2008 12:13 pm Reply with quote
Frequent contributor Joined: 22 Mar 2008 Posts: 911
Usual question R2, will this work on 5.15.x f/w (5810)? Cheers.

_________________
TF5810, TS On, F/W: MS6 Recommended F/W 12/9/2009 -Xl+UUuVdZ
TAPs: PcControl B1.3; EIT Sub (Game) v0.6; EPG2MEI v0.96; Extend v1.7; Font Manager 1.0d; SecCache (UK) v0.4; MyInfo B5.6; MyStuff 6.5;
Sig generated by MyInfo on 19/7/13
?
View user's profile Send private message

Display posts from previous:  

All times are GMT + 1 Hour
Page 3 of 4
Goto page Previous  1, 2, 3, 4  Next

Jump to:  

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum